Physicians, both directly through services they provide and indirectly through the services that they order, shape the vast majority of health care utilization in the U.S. There has long been interest in the determinants of physician decision-making because optimizing physician decisions has the potential to improve the quality of U.S. health care while also controlling costs by eliminating the use of unnecessary services. One influence on physician decision-making that has largely been ignored, however, has been physicians' position in the larger networks of doctors. In this setting, networks refer to informal networks of primary care and specialist physicians who share patients and information rather than to formal networks of all of the physicians formally affiliated with a health plan or hospital. In practice, each physician, particularly each primary care physician, develops distinct sets of specialist physicians with whom they share patients and information. By virtue of sharing patients, these networks can be defined empirically. These networks of physicians in turn have the potential to influence an individual physician's decision making, and such influences may vary depending on the structure of the network and/or occurrences within the network. In this project, we aim to improve our understanding of this phenomenon. We also aim to explore the contribution of network properties to observed variations in spending and treatment patterns within the Medicare program. Our proposal has the following specific aims: (1) to create and describe a dataset about physician networks using comprehensive national data from the Medicare program regarding over 10 million people and up to 150,000 doctors; (2) to identify the characteristics of different types of hospitals and regional hospital markets that are associated with physician networks and to examine the extent to which formal organizations influence the formation of informal networks; (3) to examine the characteristics of individual primary care and specialist physicians that are associated with properties of physician-centered and geographically delimited physician networks; and (4) to evaluate the contribution of physician networks to variations in health spending and the use of particular health care services (e.g., hospice care) according to geographic regions. The role of networks and the extent to which network connections influence the clinical decisions made by individual physicians is unknown. A better understanding of this complex phenomenon will improve our ability to influence physician decision making to improve quality and decrease costs and provide novel insights into the role that provider connections play in the delivery of health care.